In a wireless network (e.g., mobile core network), a typical Quality of Service (QoS) approach is to provide “best effort” when servicing applications. However, as applications or services utilized in the wireless network become more prevalent and more data intensive (e.g., video, multimedia), providing QoS for these applications or services can become more challenging. Further, a QoS of “best effort” is not desirable and/or sufficient for many new applications or services. Traditional QoS approaches are not sufficient due to, for instance, large uncontrollable variations in radio frequency (RF) link conditions (and resulting available bandwidth).
Conventionally, rate adaptation (e.g., coding rate, which translates to a corresponding bandwidth) is used to provide decent QoS with efficient resource utilization for voice calls in wireless systems (e.g., cellular systems); however, rate adaptation based on network conditions is generally not used for video or other applications or services, which are highly sensitive to QoS conditions such as delay/latency, jitter, packet loss, etc. Rate adaptation can be performed at the source (e.g., a video server) of the application or service, but the associated feedback (FB) loop from the radio link to the source can be very long since the source may often be deep in the network (e.g., deep in the wire-line portion of the network), not close to the edge of the wireless portion of the network, and the amount of time for FB communication between the source and radio link increases with an increase in distance between the source and radio link. As a result, often the FB adaptation delay can be too large to be effective for fast-changing conditions such as in the RF link (e.g., the rate coding can change, but by the time the change is performed at the other end of the FB loop (e.g., at the source), the RF conditions may have already changed again, so as to make the adapted rate inappropriate by the time it is changed). Bringing a full-bandwidth/full rate/full resolution data stream all the way from the source to the wireless edge would allow rate adaptation closer to the mobile communication device, which would allow sufficiently fast adaptation, but at large scale would consume an undesirable (e.g., enormous) amount of bandwidth resources in the wire-line portion of the network. It is desirable to provide sufficiently fast adaptation time to change conditions in the wireless portion of the network without consuming an undesirable (e.g., enormous) amount of wire-line bandwidth resources.